Yeah it's frustrating to see that being scored as Informative. AI generated code isn't perfect but it's clearly faster/better than googling the forums and tutorials.

Yeah gpt-o1 code is really really good, I only have to fix about 10% of it. I keep meeting people who disagree but, it works for me, I have written dozens and dozens of programs in the last two months, probably 5x faster.

Not true, lol, never heard that argument be made by any computer scientist. All of the good AI training libraries will award negative score for directly copying the training data.

Hard disagree.

I'm a AAA graphics programmer, I'm at the top of my game, 15 years experience on some big titles, a graphics engine that ships billions of dollars of games. When you use it properly, GPT absolutely rocks at programming for real world huge scale problems. You can quote me: it's insane to lowball GPT's ability. GPT knows intricate details about how to handle complex high performance code.

If you are not prepared and it gives you a hallucination, or you try to let it lead, then yes, it can't code. One example - it can write itself in a corner where it needs a function that doesn't exist. Next time you query - just tell it that it is having a problem because it keeps trying to use the same non-existant function. Problem solved.

Yes, it's scope is limited - can't handle more than about 200 lines of code at once. I just break up my ideas into pseudocode, capture the dependencies, and write them into the prompts to generate the functions.

Yes it makes simple mistakes. I just correct it and move on.

If you work through the problem with GPT methodically, it will solve it, absolutely, with enough retries, 95% of the time.

This is an absolutely insane productivity boost for me, I'd estimate 5x or more.

"Any program written to stop AI harming humans and destroying the world, for example, may reach a conclusion (and halt) or not -- it's mathematically impossible for us to be absolutely sure either way, which means it's not containable."

This is mumbo jumbo.

Conceptually, you need a sensor that detects human in harm's way. Then the AI should be disabled from acting when that sensor is true.

If you want something more than that, fine - but don't act like it can't be done. You're the human, there's nothing you "can't" make it do.

I wouldn't personally spend money on ECC RAM. I've worked on many dozens of computers and hundreds of code bases as a programmer and I have never pinpointed RAM corruption as a noteworthy failure point. There are a lot of things to worry about in life, like failing hard drives. RAM corruption at most requires a reboot.

It's power without responsibility. This is what Malcolm tried to warn Hammond about - you work off the backs of giants, but you never earned the knowledge.

The consequences are probably less dire than a failing T-rex enclosure though.

None of our remote vehicles EVER come back. You try piloting remotely at Astronomical scales.

AAA Game programmer here. I'm disappointed to see so many dismissive and rude responses.

I would estimate the speedup as at least 100x (just an educated ballpark). I choose this number because, the speedup for a single shader would be well over 1000x, but games don't use just one shader.

Game engines use different shader programs, from dozens to thousands, in order to draw different techniques. The first constraint would be that you'd need to choose which shaders you want. Shaders are often *a little bit* generic, in that, many, many of them are doing essentially the same things but slightly differently.

First, you would convert all of your shaders into a generic, branching architecture, an "uber-shader" (this is the actual jargon).

If the artists agreed that you have ensapsulated all novel rendering techniques that they care about, then you could start building an ASIC architecture to replace those shaders.

But development cost would be high and you would not be able to support every game.

Yes, machine learning *is* unpredictable. But the point is that the existing dataset is so dominated by class division, you've gotta predict that there's a reasonable chance the machine will generate a model that reinforces the old class divisions, even if it has to make tenuous links.

Hiring should be *mostly* based on merit, but we also have a chance to improve the future by taking a conscious hand in deviating from the pure numbers. Ie, the human factor.

The difference is that events A and B *did* actually have an objective ordering, it was simply that the observer's measurements (which relied on photons reflected off the event) has distortion. But you're partially right about subjectivity of the observer.

The waveform collapse *doesn't care* about whether you're "microscopic" or "macroscopic" or "two different people" but it does care if 2 particles interact and become entangled, in which case, they have to be consistent.

When two physicists, Alice and Bob compare results of the same quantum experiment, *every particle that interacts in their bodies and environments* will collapse its waveform, resulting in one consistent history for the whole room without paradox. The waveform collapse will propagate through every quantum element it interacts with (via light bounces, strong or weak forces, any interaction). The "waveform collapse" just defines which reality for particle A is connected to which reality for particle B if they're going to interact.

Until Alice and Bob affect each other, for each Alice, there's an infinite Bobs (even impossible Bobs), and vice versa. But when any particles from Bob interact with particles from Alice, those particles become entangled, and the waveform collapses (for that particle). The waveform collapse propagates to any other particle affected by the original particle. So all the particles that affect each other are in the same, consistent universe.

Now if Charlie has Alice and Bob in a box, then for Charlie, there are STILL infinite sets of entangled "Alice and Bob" pairs in superposition. Charlie won't know which Alice and Bob he's connected to until *he* measures them.

Also notice, this is the same conclusion the layman can make about subjective reality. A person can't know what's inside a lead-lined box. There's no objective different between saying "I can't know what's in the box until I look" and "All infinite possibilities are in the box, and when I open it, I will find 1 of those realities."

The fan boys are just that... boys. NMS is for men and women who love the idea of space exploration.

As an adult (35 year-old) AAA developer who has worked on some of the best rated space games of all time... Let me tell you, No Man's Sky is a treasure. It's a game-making achievement. I am a game dev, and I studied NMS deeply. These tools and tricks have existed for only about 15 years and never in the same product before to such a high quality.

There's not very much that hits my quality bar, but I put at least 40 hours in. That means, it's an immense accomplishment for Hello games.

The juvenile can cry all day about 'promises'. It doesn't make a difference. What was delivered is pure gold.